Estimation of Solubility Parameters for Some Olefin Polymers and

Calculation of Solubility Parameters by Inverse Gas Chromatography. Gareth J. Price. 1989,48-58. Abstract | PDF | PDF w/ Links. Cover Image ...
0 downloads 0 Views 601KB Size
Solubility Parameters for Some Olefin Polymers 1163

Vol. 12, No. 6, November-December 1979 (18) Y. Mitsuda and J. D. Ferry, Kobunshi Ronbunshu, 31, 135 (1974). (19) . , J. W. M. Noord'ermeer. 0. Kramer. F. H. M. Nestler. J. L. Schrag, and J. DI. Ferry, Macromolecules,8, 539 (1975). (20) (a) B. H. Zimm and R. W. Kilb, J . Polym. Sci., 37,19 (1959);

(b) K. Osaki, Y. Iklitsuda, J. L. Schrag,and J. D. Ferry, Trans.

SOC.Rheol., 18, :395 (1974).

(21) . , K. Osaki and J. L. Schrae. -, J . Polvm. Sci.. Polvm. Phvs. Ed.. 11, 549 (1973). (22) T. Masuda, Y. Ohta, and S. Onogi, Macromolecules, 4, 763 (19711. -, (23) A. L. Soli, Ph.D. Thesis, University of Wisconsin, Madison, WI, 1978. (24) D. J. Massa, Ph.!D. Thesis, University of Wisconsin, Madison, WI, 1970. (25) D. J. Massa and J. L. Schrag, J . Polym. Sci., Part A-2, 10, 71 (1972). (26) M. H. Birnboim, J. S. Burke, and R. L. Anderson, I

(27) (28) (29)

I

\ - -

(30) (31)

"Proceedings of the Fifth International Congress on Rheology", Kyoto, S. Onogi, Ed., University Park Press, Baltimore, MD, 1969, p 409. C. Sadron, J. Phys. Radium, 9, 381 (1938). J. L. Schrag, Ph.D. Thesis, Oklahoma State University, Stillwater, OK, 1967. J. D. Ferry, "Viscoelastic Properties of Polymers", 2nd ed., Wiley, New York, 1970. G. Bi Thurston and J. L. Schrag, J. Polym. Sci., Part A-2, 6, 1331 (1968). J. L. Schrag and J. D. Ferry, Faraday Symp. Chem. Soc., No.

6. 182- (19721. -, ~ ~~-, (32) N. Hadjichristidis, A. Guyot, and L. J. Fetters, Macromolecules, 11, 668 (1978). (33) J. E. L. Roovers and S. Bywater, Macromolecules, 5, 385 (1972). (34) M. Nagasawa, private communication.

Estimation of Solubility Parameters for Some Olefin Polymers and Copolymers by Inverse Gas Chromatography K. Ito' and J. E. Guillet* Department of Chemistry, University of Toronto, Toronto, Canada M5S 1Al Received May 22, 1979 ABSTRACT: Flory-Huggins x parameters have been determined by gas chromatography at 30 O C for ten standard hydrocarbons in ethylene (40%)-propylene (60%) copolymer, cis-polyisoprene, and amorphous polypropylene to estimate the polymer solubility parameter (6,") by the method developed by DiPaola-Baranyi and Guillet. The results are m follows: 6", = 7.70 h 0.11 for ethylene-propylene, 7.96 h 0.10 for cis-polyisoprene, and 7.67 f 0.16 for polypropylene. The parameters for ethylenepropylene were also estimated by extrapolation from the higher temperature data (63-83 "C) and found to be in good agreement with those estimated directly. The solubility parameter concept has been used extensively in practical applications of polymers. DiPaolaBaranyi and GuilleP have recently shown that inverse gas chromatography, using a polymer as the stationary phase, can be a simple and convenient method for estimating solubility parameters for polymers. The method is based on the principle that the Flory-Huggins x parameter can be readily determined from retention data on various small molecule probes and that x can be related to solubility parameters by Hildebrand-Scatchard theory4 combined with Flory t h e ~ r yas , ~follows:

x where

= (Vi/RT)(a, - 82)'

x has free-energy characteristics, i.e., x = XH + XS

(1) (2)

and V1 is the molar volume of the probe used, 61 is the solubility parameter of the probe, and & is the solubility parameter of the polymer. In the usual experiment, x is determined under conditions approximating infinite dilution of the probo in the polymer, and hence the value of a2 is more correctly designated a2", since it is also an infinite dilution quantity. For reasons outlined previously,2 this quantity may have more fundamental significance than the value of d z measured a t finite solute concentrations by classical methods such as swelling or solubility measurements. Furthermore, a t the high dilution of the experiment, it is possible that the assumptions of regular solution behavior, inherent in the Hildebrand-Scatchard treatment, are more closely fulfilled. T h e method has been successfully applied to estimate consistent b2" values for polystyrene at 193 "C, poly(methy1 0024-9297/79/2212-1163$01.00/0

acrylate) a t 100 "C, and poly(viny1 acetate) at 135 "C, using a variety of standard hydrocarbon solutes of different a1 values. bzm was also estimated a t 25 "C, using x parameters computed by extrapolation of the higher temperature data according to (3) x = a P/T where the constants a and fi should have characteristics of entropy and enthalpy, respectively. This empirical equation is usually valid only for relatively small ranges of temperature; however, the values thus obtained were shown to agree well with literature data, supporting the idea that the derived parameters might be practically useful in predicting polymer solubility even a t finite concentrations. The present work extends the method to polymers including ethylene-propylene copolymer (EPR), cis-polyisoprene (PIP), and amorphous polypropylene (PP). Since these polymers have very low glass-transition temperatures (-20 "C or lower), the GLC data can be obtained close t o room temperature, so that their 62- values can be determined directly and compared with those estimated by the extrapolation method. It will also be shown that the x parameters are more properly expressed by the relation derived by Huggins' and by Scott and Magat?

+

w

x = ( V l / R T ) ( ~ ,- + Y (4) where y is an entropy correction term which is often considered to be a constant near 0.3 for many polymer^.^^^ Experimental Section Ethylene-propylene rubber was Exxon Vistalon 404 with a composition of 40% ethylene and 60% propylene. cis-Poly0 1979 American Chemical Society

Macromolecules

1164 Ito, Guillet Table I Column Parameters

~

packing and column .~

a

loading,

polymer ethylene-propylene rubber

coating solvent xylene

cis-polyisoprene

benzene

8.25

polypropylene

benzene

7.29

_--

polymer mass, g 0.412 0.0313 0.329 0.0826 0.373 0.08 26

%a

8.34

length, in. 18 12 18 4 18 12

o.d., in. 114 118 114 114 114 118

Inert support was Chromosorb G, AW, DMCS, 70-80 mesh. Table I1 Thermodynamic Parameters of Hydrocarbons in Ethylene-Propylene Rubber at 7 3 "C solute

Vl cm3/mol

2,2,4-trimethylpentane n-hexane n-octane n-decane cyclohexane tert-butylbenzene ethylbenzene benzene

176.2 141.4 173.4 206.3 115.6 163.0 129.5 95.1

AH,,, kcal/mola

AH,",^

9

obsd 7.81 i 6.99 c 9.20 * 11.41 c 7.36 f 10.79 i 9.47 * 7.53 i

calcd 0.04 0.18 0.04 0.07 0.08 0.07 0.03 0.02

7.78 6.83 9.10 11.49 7.29 10.81 9.43 7.45

6,

6.36 6.68 7.01 7.21 7.60 7.87 8.24 8.49

(al/wl)" 4.9 5.5 4.7 4.4 3.9 3.7 3.8 4.3

The error limits quoted are variances ( l o ) based on least-squares analysis of the data. AH," = R a In ( a l / w l ) " / a ( l / T ) .

kcalimol 0.31 c 0.02 0.25 i 0.09 0.33 c 0.03 0.24 c 0.05 0.22 i 0.05 0.46 i 0.04 0.51 i 0.02 0.57 * 0.02

X

0.33 0.38 0.30 0.28 0.22 0.27 0.31 0.44

In ( a l / w l ) " = In (273.16Ri

p I oVgil/ll)- plo(Bll- Vl)/RT.

Table I11 Temperature Dependence of the x Parameter of Hydrocarbons in Ethylene-Propylene Rubber intercept

a

x = CY

solute

ff

slope P

correl coeff

2,2,4-trimethylpentane n-hexane n-octane n-decane cyclohexane tert-butylbenzene ethylbenzene benzene

-0.413 -0.389 -0.466 -0.299 -0.418 -0.577 -0.670 -0.703

256 26 6 266 203 222 295 338 3 94

0.9982 0.9350 0.9950 0.9726 0.9754 0.9899 0.9988 0.9996

obsd 0.42 0.46 0.42 0.34 0.33 0.37 0.45 0.60

+ PIT

isoprene (Mn 126000) was obtained from Polysar, and amorphous polypropylene was obtained from Eastman (Epolene D-10). These polymers were completely soluble in the solvents indicated in Table I and were used as supplied. Column preparations and collection of retention data were carried out as b e f ~ r e .Two ~~~ columns of different size containing low and high polymer loadings were prepared for each polymer (see Table I) and used with ten solutes, including 2-methylbutane (bp 28 "C) and tert-butylbenzene (bp 169 "C), within an accessible range of retention time from 1 to 40 min a t a flow rate of 30 mL/min. The retention volume was confirmed to be independent of solute sample size in all cases studied, and reproducibility was good (within * 5 % ) between columns with different loadings. Data reduction was followed as described in previous publication~.%~ The following relations were used to obtain the specific retention volume V, (corrected to 0 "C) and the x parameter: (5)

x

x a t 30 "C calcd" 0.43 0.49 0.41 0.37 0.31 0.39 0.45 0.60

= In

(

273.16Rv2 PlOV,V1

)

-

1 - pl0(Bll - V1)/RT

(pl0),second virial coefficient (Bll), and the molar volume (VI), were obtained or computed from literature sources10 as b e f ~ r e . ~ ~ ~ The polymer specific volumes (uz) for EPR," PIP," and PP13were computed from literature data.

Results and Discussion Various thermodynamic parameters, including weight fraction activity coefficient (al/wl)", partial molar heat of mixing AH1" at infinite dilution, and the Flory-Huggins parameter x,were first determined for eight hydrocarbons i n EPR b y measuring the retention volumes over the temperature range 63-83 "C. Table I1 summarizes th_ese data at the midpoint, 73 "C. The values of (al/wl)", AH1", and x a r e all of the magnitude expected for t h e polymer-solvent case. T h e y a p p e a r t o have their m i n i m u m (corresponding to the largest solubility) around the solutes n-octane (6, = 7.01), n-decane (6, = 7.21), or cyclohexane (6, = 7.60). The polymer solubility parameter a2 for EPR was then evaluated from the relation derived from eq 1:

(6)

where t R is the net retention time, F is the flow rate measured with a soap bubble flow meter at ambient temperature TI (K), w is the mass of polymer in the column, Piand Poare the column inlet and outlet pressure, and P, is the water vapor pressure at TI.The solute (probe) parameters, including the vapor pressure

(7)

(Sl2/RT) - (x/V,) against 6 is shown in Figure la. The linear correlation is excellent, and a least-squares analysis gave 62 = 7.18 f 0.11 and 7.26 f 0.12 from the slope and t h e intercept, respectively, at 7 3 "C.

T h e plot of

Solubility Parameters for Some Olefin Polymers 1165

Vol. 12, No. 6, November-December 1979

Table IV V,, 6 ,, and x Values for Hydrocarbons in Ethylene-Propylene Rubber, cis-Polyisoprene, and Polypropylene at 30 " C X

EPR solute

VI

2-methylbu ta.ne 2,2,4-trimethylpentane n-pentane n-hexane n-octane n-decane cyclohexane tert-butylbenzene ethylbenzene benzene a

Equation 4 ; 6

118.4 167.1 117.0 132.5 164.5 196.9 109.4 156.4 123.7 90.0

:= 7.70, y = 0.33.

PIP

PP

6,

obsd

calcda

obsd

calcdb

obsd

calcdC

6.63 6.83 6.95 7.20 7.49 7.67 8.15 8.29 8.74 9.08

0.59 0.42 0.52 0.46 0.42 0.34 0.33 0.38 0.45 0.60

0.56 0.48 0.44 0.39 0.34 0.33 0.37 0.42 0.55 0.61

0.57 0.46 0.52 0.50 0.46 0.43 0.29 0.22 0.29 0.41

0.62 0.62 0.47 0.40 0.33 0.30 0.28 0.30 0.40 0.46

0.91 0.58 0.64 0.48 0.55 0.29 0.44 0.61 0.72 0.74

0.68 0.67 0.57 0.57 0.48 0.47 0.51 0.57 0.71 0.77

Equation 4 ; 6

= 7.96, y = 0.27.

Equation 4;6

=

7.67, y = 0.47.

Table V Estimation of 6 and y Parameters for Ethylene-Propylene Rubber, cis-Polyisoprene, and Polypropylene ~

6

system

slope

EPR, 73 "IC: EPR, 30 "C PIP, 30 " C PP, 30 " C

7.18 7.70 7.96 7.67

-i

0.11 0.11

i.

0.16

i

* 0.10

from eq 7 intercept 7.26 7.80 8.03 7.81

i i

t i

0.12 0.10 0.09 0.16

?/VI from intercept of eq ga 0.0016 0.0026 0.0019 0.0035

y from

eq 1 1 ~ 0.24 0.35 0.26 0.48

a y/V, = I-intercept1 of ( 6 l z / R T - x / V , ) vs. 6 , plot - 6,'/RT. y = [y/V1]aboveVl(av), where of V, of the solutes used. Average with standard error for y = X o b d - (V,/RT)(6, - 6 *)'.

The x values obtained over the temperature range 63-83 "C were analyzed according to eq 3 to estimate the constants CY and @. The results are given in Table 111, where the corresponding correlation coefficient shows, in general, a good linear correlation between x and 1/T. The values of CY and P so determined were then used to calculate x values at 30 "C by extrapolation. Table I11 includes these x values as well as those determined directly from experiments a t 30 "C. The agreement is quite good, which lends support to the method used previously to calculate x and 6 values a t 30 O C from data obtained a t higher temperatures.' Since the molar volume V1 and the latent heat of vaporization AH, of the standard hydrocarbon solutes is readily available or may be computed from the literature,'O the solubility parameter 6' of the solute probe can also be easily estimated a t any desired temperature from the definition 61 = [(AH" - RT)/Vl]"2

(8)

AHv, the heat of vaporization of the probe molecule, can be obtained directly from chromatographic data from the relation AH, = ARlm- AH,

(9)

where A",is the experimental heat of solution and film is the partial molar heat of mixing. These quantities are determined from the variation of the retention volume and the activity coefficient, respectively, with temperature.' AHv can also be computed from the relation AH, = MW(d - e t ) (10) where t is the temperature in "C and d and e are constants given in ref 9. Table I1 includes, for comparison, the AH, values calculated from eq 10, which show very good agreement with those determined from the retention data. Figure l b shows the linear plots of (S12/RT) - (x/V1) against 6', using the x values extrapolated to 30 "C. The line is indistinguishable from the extrapolated values

eq 4c 0.21 0.33 0.27 0.47

i i

t t

0.07 0.07 0.10

0.11

is the average

6 Figure 1. Estimation of solubility parameter for ethylene-propylene rubbers: (a) 73 "C, G,(slope) = 7.18 f 0.11, &(intercept) = 7.26 f 0.12; (b) extrapolation to 25 "C, d,(slope) = 7.61 f 0.11, ?i,(intercept) = 7.69 i 0.12; (c) direct measurement at 30 "C, bz(slope) = 7.70 f 0.11, &(intercept) = 7.80 f 0.10.

shown in Figure l a , and the derived a2 values by leastsquares analysis with eq 7 are 7.61 f 0.11 and 7.69 f 0.12 from the slope and the intercept, respectively. The corresponding values, using the x data observed directly (Figure IC), are 7.70 f 0.11 and 7.80 f 0.10, in good agreement with the above, again supporting the previous extrapolation procedure for estimating 62 a t room temperature. Table IV summarizes the x values determined a t 30 "C for ten hydrocarbon solutes in EPR, PIP, and PP together with the V1 and 6' values. x values in all systems are smaller than unity, as expected for polymer-solvent systems. It appears from comparison of the x values that the hydrocarbons in general are slightly less soluble in PP than in P I P or E P R and that PIP tends to dissolve aromatic probes more readily than E P R and PP. Plots of (S12/RT) - (x/V1) against 61 are shown in Figures ICand 2, and the & values determined by a least-

1166 Ito, Guillet

Macromolecules

6

7

8

9

*I

Figure 2. Estimation of solubility parameter for: (a) cis-polyisoprene at 30 "c,&(slope) = 7.96 i 0.10, &(intercept) = 8.03 f 0.09; (b) polypropylene at 30 "C, &k+lope)= 7.67 f 0.16, &(intercept) = 7.81 f 0.16. Table VI

x Parameters for Hydrocarbons in Ethylene-Propylene Rubber at 30 " C Xcalcd

Y

(=x -

Xcalcd

solute

Xobsd

from eq la

2-methylbutane 2,2,4-trimethylpentane n-pentane n-hexane n-octane n-decane cyclohexane tert-butylbenzene ethylbenzene benzene

0.59 0.42

0.23 0.21

0.36 0.21

0.56 0.48

0.52 0.46 0.42 0.34 0.33 0.37 0.45 0.60

0.11 0.06 0.01 0.00 0.04 0.09 0.22 0.28

0.41 0.40 0.41 0.34 0.29 0.29 0.23 0.32

0.44 0.39 0.34 0.33 0.37 0.42 0.55 0.61

a With 6 = 1.70. y = xoba With 6 = 7.70 and y = 0.33.

-

(V,/RT)(6

from

-

6

2)2.

squares analysis are summarized in Table V. It should be noted that, in applying eq 7, the 6zvalues derived from the intercept were slightly but consistently higher than those from the slope and moreover that in spite of the excellent linearity of the plots in Figures 1 and 2, the x values recalculated from eq 1,using 62 calculated from the slope, were always significantly lower than those observed, as exemplified in Table VI for EPR, 30 "C system. In order to resolve this discrepancy, eq 4 derived by Huggins7 and by Scott and Magats was used in place of eq 1. This can be rewritten as

RT

Vi

Therefore, the plot of (b12/RT) - (x/VI) against 61should also give 2a2/RT as the slope but give (6,ZlRT)+ (y/Vl) instead of 6:/RT as the intercept. Fortunately, the contribution of the term y/V1 to the intercept may be estimated to be only 3% or less,gso that in this type of plot very similar values of 62 can be obtained both from the slope and the intercept, although those from the slope should be taken as more correct. In fact, in all of the present systems (Table V), the estimated value of y/V1 from the observed intercept was found to be in the range from 0.0016 to 0.0035, which corresponds to only 2 to 3% error in the intercept. Although the contribution of y/V1 in the plots according to eq 11 can be relatively small, the

values of y,estimated from these r / V l values, are in a range of 0.24 to 0.48 (Table V), which will make an important contribution to the predicted value of x parameters, particularly when x is small, as in the present case. For example, Table VI summarizes the experimental x parameters obtained for various hydrocarbon probes in EPR a t 30 "C along with the x calculated from eq 4, using = 7.70. The calculated x values are always significantly lower than the experimental ones, and this difference can be equated to the value of y in eq 4. The values of y range from 0.21 to 0.41 and average 0.33, which is definitely outside the experimental error in both x and &.. Using eq 4 with the same value of a2 and y = 0.33, the average value in these experiments, we obtained a value of x (last column in Table VI) much closer than that from eq 1, probably representing a reasonable approximation to the correct value. However, it does appear that there are consistent differences between the average value of y with different classes of probe molecules (yav= 0.39 for linear alkanes vs. 0.27 for aromatic probes), and this may relate to a real difference in this excess entropy term. The values of aZm from the slope and intercept and the corresponding average value of y are listed in Table V for EPR a t 30 and 73 O C for PIP and a t 30 "C for PP. There is considerable variation in y for the three different polymers, but again the average is 0.33. From these data one can conclude that the use of eq 4 with y = 0.33 should give a reasonable approximation in calculating x parameters from the solubility parameters of the polymer and probe. Furthermore, if eq 1 is used to estimate aZmvalues from GC data, only the value obtained from the slope should be used, since it does not depend on the value of y, which may vary from case to case. From these studies it is concluded that the inverse GC procedure outlined in this and previous paper^^,^ does indeed give a rapid and useful method of estimating the solubility parameter of a polymer, aZm, at infinite dilution of the probe molecule. Furthermore, extrapolation of the observed x values to 30 "C appears to be justified in order to calculate tjZmat temperatures comparable to those a t which the solubility parameter is usually measured and reported. Although eq 4 clearly gives a more exact representation of the experimental data than eq 1,the value of should not be affected, provided that it is evaluated from the slopes of the relationships given by eq 7 and 11 rather than the intercepts. Except for amorphous polypropylene, the cSZm values obtained in these experiments are very close to those of the conventional a2 values tabulated in the literature. For example, the value for P I P of aZm = 7.96 f 0.10 is very close to that for natural rubber at 25 "C (7.9 to 8.3).'* Similarly, current values of 7.70 f 0.11 for EPR and 7.67 f 0.16 for amorphous PP are very close to that for polyethylene (7.7 to 8.8, mostly near 7.9).'* Crystalline polypropylene is usually reported to have a solubility parameter of 9.2 to 9.4, which is considerably higher than the value of bZmdetermined in these laboratories for the amorphous polymer. This difference is greater than one would expect from density differences alone and may be related to the experimental problems involved in determining solubility parameters for highly crystalline polymers. The current value indeed seems to be more consistent with what would be expected for a branched alkane polymer without strong crystalline interactions. Acknowledgment. The authors wish to acknowledge the generous financial support of the National Research Council of Canada and helpful discussions with Professor J. Klein of the Technical University of Braunschweig.

Poly(pheny1 vinyl ketone) 1167

Vol. 12, No. 6, Nouember-December 1979

References and Notes (1) Nagoya University, Department of Synthetic Chemistry, Forocho, Chikusa-ku, Nagoya, Japan. (2) G. DiPaola-Baranyiand J. E. Guillet, Macromolecules, 11,228 (1978). (3) G. DiPaola-Baranyi,J. E. Guillet, J. Klein, and H.-E. Jeberien, J . Chromatogr., 166, 349 (1978). (4) J. H. Hildebrand and R. L. Scott, “The Solubility of Nonelectrolytes”, 3rd ed., Reinhold, New York, 1950. ( 5 ) P. J. Flory, “Principles of Polymer Chemistry”, Cornell University Press, Ithaca, N. Y., 1953. (6) D. Patterson, Rubber Chem. Technol., 40, 1 (1967).

(7) M. L. Huggins, Ann. N.Y. Acad. Sci., 43, l(1942). (8) R. L. Scott and M. Magat, J . Chem. Phys., 13, 172 (1945);J. Polym. Sci., 4, 555 (1949). (9) G. M. Bristow and W. F. Watson, Trans Faraday Soc., 54, 1731 (1958); 54, 1742 (1958). (10) D. R. Dreisbach, Adv. Chem. Ser., No. 15 (1955);No.22 (1959); No.29 (1961). (11) “Encyclopedia of Polymer Science and Technology”, Vol. 6, Wiley-Interscience, New York, 1971. (12) J. Brandrup and E. H. Immergut, Ed., “Polymer Handbook”, Interscience, New York, 1966. (13) S. Newman, J . Polym. Sci., 47, 111 (1960).

Triplet Energy Migration in Polymer Photochemistry. A Model for the Photodegradation of Poly(pheny1 vinyl ketone) in Solution M. V. Encinas, K. Funabashi, and J. C. Scaiano* Radiation Laboratory,’ University of Notre Dame, Notre Dame, Indiana 46556 Received May 31, 1979

ABSTRACT: Plots of the progress of the degradation of poly(pheny1 vinyl ketone) as a function of irradiation time are nonlinear; the effect is the result of triplet quenching by unsaturated end groups generated in the reaction. The process is intramolecular, and from a study of polymers of different initial chain length and application of random walk theory for a one-dimensionallattice it is possible to determine the frequency of triplet energy hopping; the value obtained was -1 X 10l2s-l at 30 “C in benzene. Poly(pheny1vinyl ketone) has a critical chain length of 750 units, this being the average number of chromophores “visited’by the excitation during the triplet lifetime.

from the triplet m a n i f ~ l d . ~ ~ ~ ~ ~ The transfer of electronic energy is a process of importance in numerous s y s t e r n ~ . ~The - ~ migration of electronic energy between chromophores of the same type has received attention in solution and in the solid ~ t a t e . ~ - ~ While singlet energy transfer can occur by several mechanisms, triplet-triplet transfer can only occur by the exchange mechanism, which requires the donor and acceptor A groups to be close enough to allow significant overlap of their electron cloud^.^,^ Energy migration between identical groups is of importance in crystals, glasses, and polymer systems in general. Transfers between truly identical chromophores Ph Ph are virtually impossible to study because the energy B transfer process does not lead to an easily observable phenomenon, and most studies make use of small differWhen the degradation process is monitored as a function ences between the donor and acceptor molecule. For exof the irradiation time, the corresponding plot shows sigample, in the case of polymers, the donor and acceptor may nificant curvature (Figure 1). This curvature becomes differ only in that one of them will be closer to the end quite evident a t extremely low conversions, that is, under of the chain, while they are identical in all other respects. conditions where the number of chromophores which have Energy migration of this type is often responsible for the participated in reaction 1 is only a small fraction of the occurrence of chemical or physical processes a t a considtotal number of chromophores, typically -0.1%. Quite erable distance from the original excitation site. This clearly, the effect cannot be attributed to the depletion of phenomenon has important practical implications in the reactive sites, nor is it the result of UV screening by field of photostabilizers, some of which work on the prinproducts or impurities in the polymer (see Results Section). ciple of excitation migration toward a suitable energy This unusual effect attracted our attention to the problem, sink;1° processes of this type can only be of importance if which we propose results from intramolecular triplet migration “funnels” the energy toward the sink. quenching by a,@-unsaturatedcarbonyl chromophores of The photochemistry of polymers in the presence of the type shown in fragment B in reaction 1. These groups “impurity” traps has been the subject of several studies behave as extremely efficient quenchers; we believe that in films and gla~ses,*J-’~ but little is known about the the high efficiency is the result of fast energy migration, behavior in solution.lgZ6 which eventually funnels the energy to the end chromoPoly(pheny1vinyl ketone), PPVK, has been the subject phores in the polymer, effectively leading to quenching by of numerous photochemical s t ~ d i e ~The . ~poly~ ~ ~remote ~ ~ energy ~ ~ ~sinks. ~ ~The ~ importance - ~ ~ of end groups in mer undergoes cleavage via the Norrish type I1 reaction controlling migration processes had also been recognized

v w

w

0024-9297/79/2212-1167$01.00/0 0 1979 American Chemical Society